The current status of instructional design theories in relation to today's authoring systems

It is of course very difficult to accurately project important characteristics of the future state of any rapidly evolving field, and the field of authoring systems for computer-assisted instruction (CAI) is no exception. However, strong trends in evolving CAI systems of today would seem to indicate some important characteristics of the software systems underlying tomorrow’s automated instructional systems. Under the premise that form follows function, this paper will explore the functional impetus of some of the most important trends extant in terms of their roots in a powerful, emerging technology of instructional design. This technology promises to complement the advancing hardware and software technologies, by supplying that sine qua non of successful instructional materials. It is inevitable that systems of the future will make (and indeed are making) significant accommodations of this vital technology as it matures, and therefore some important characteristics of tomorrow’s systems can be predicated upon what is already known of this technology of instructional design. Introduction One of the most important influences of the new instructional technology stems from the fact that it is a technology. Rooted in a growing body of research prompted by the empirical observations of a large body of practitioners, defined by an increasingly rigorous calculus of practice, and validated by an impressive range and volume of application, the technology has matured to the stage where it is sufficiently well defined to support instructional engineering applications by non-theoretically inclined users. *This paper was first published in two parts in Interactive Learning International Vol. 2, nos 3 & 4, 23–26 and 29–33, by John Wiley & Sons Ltd. British Journal of Educational Technology Vol 39 No 2 2008 251–267 doi:10.1111/j.1467-8535.2008.00815.x © 2008 The Author. Journal compilation © 2008 British Educational Communications and Technology Agency. Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK The implications of this degree of definition needs to be explored in terms of tomorrow’s authoring systems. One of the questions this paper will address is ‘To what extent can (or should) we build today’s instructional strategies into CBT authoring systems?’ CBT languages, as they have been developing, although they may be mechanically efficient and intellectually seductive, are essentially instructionally irrelevant. That is, although they provide the tools for defining interfaces between users (authors, students, proctors, instructors, etc.), the content and the systems capabilities, these are at best only definitions of mechanical boundary conditions (Bunderson, 1977). The languages are deficient in the metaphors of the instructional technology; they imply no necessary considerations of primitive content structures, strategy definition, or locus of control. Most are written on a level of discourse far below that which would be desirable (Pask, 1969). In short, whereas they may serve to communicate to the system ‘how’ to do something, they offer little help to the instructional designer in terms of the more important question, namely ‘what should be done’ instructionally. Therefore one important evolutionary step in the CAI systems of tomorrow will be increasingly sophisticated ‘authoring systems’ as opposed to ‘authoring languages’. To be sure, these ‘authoring systems’ will perforce include authoring languages as an important component, but more significantly, they will include software and data structures for cueing, prompting and tutoring the instructional designer/developer in ‘what’ he should be doing instructionally, based on the analytical and prescriptive models implied by the design technology. This facet of software development is being actively pursued by several different groups. Current programs, approaches and progress will be discussed in the body of the paper, as will a more detailed description of the types of functions a mature authoring system should include. Another important influence on the software architecture of coming CAI systems is the trend towards well-defined componentised content structures. The efficiency and economy of authoring which can be realised with componentised content (eg, content structures which are discrete and well defined, such as rules, definitions, examples practice items, introductions, summaries, helps, etc) are attractive. The ‘factory approach’ to instructional content development and coding is extremely powerful in a componentised environment (Bunderson, 1973). Another strong influence on tomorrow’s systems is made possible by separation of content and strategy. Once basic instructional strategies can be built into the system to deal very directly and efficiently with the classes of instructional problems for which strategies and components have been defined (again, this would most efficiently be done at a higher level of discourse than that of most current authoring languages). VAULT, TICCIT, CDS and WISE represent beginning steps towards this desirable capability. The paper will discuss the nature of componentised content structures and the separation of content and strategy, and examples of system software accommodations, which might be implied, will be described. 252 British Journal of Educational Technology Vol 39 No 2 2008 © 2008 The Author. Journal compilation © 2008 British Educational Communications and Technology Agency. With componentised content structures and generally definable strategies, it becomes possible for tomorrow’s CAI systems to negotiate or manipulate the locus of instructional control as never before (O’Neal, 1973). Definition of the software structures needed will require development of both state and trait ‘advisors’ and models of the learner which will emerge as CAI systems mature. Some exist now, empirically determined, as the basis of powerful adaptive models (ie, the basic mathematical skills materials developed at Stanford, etc), but if any degree of learner control is to be allowed, the software structures must be capable of functioning in a communications and tutor mode, as well as the more familiar executive function, controlling student progress. Finally, the automated instructional systems of tomorrow must more effectively deal with the fact that not all instruction takes place on the system itself. The artificial lines between CAI and CMI (computer-managed instruction) will be broken down and a full set of management functions (both for managing instruction and for managing the development of instruction) will be necessary as part of any system. Implications of the emerging instructional design technology for authoring-system development will be outlined using examples from existing systems’ capabilities to provide perspective. A brief history of instructional design The history of instruction reveals three different periods or stages of development, which recapitulate the history of science: the artistic approach, the empirical approach and the systematic approach. In the artistic approach, the design, delivery and evaluation depends upon intuition and personal experience. To the extent that one person’s intuition or experience was better than another’s, better instruction ensued. This approach to instructional development, can generally be characterised as having a ‘cottage industry’ flavour. That is, when it comes to the development of a given segment of instruction the author becomes a jack of all trades. He does the analysis, design writing, sometimes the design of graphics, and test developments. He approaches each segment as a completely new task and employs an idiosyncratic or artistic approach to this ‘new’ instructional development problem. The model of ‘instructional developer as artist’ has several serious deficiencies. First, it is inefficient, in that much time is spent inventing the new approach to each segment. Secondly, art is difficult to evaluate or teach. Quality is variable, and the managers of CAI curriculum development activities find it difficult to train and maintain the required number of artists. Finally, artists are, alas, temperamental. Their interest in the new art (CAI) is often very temporary. They have difficulty tolerating extended periods of routine activity such as inputting instructional sequences. And, by definition, their output is difficult to predict or control. These factors lead to uneven rates of production, inability to meet deadlines, and several other undesirable results. They have been a major cause of the current state of the bibliography of CAI materials, which includes The current status of instructional design theories 253 © 2008 The Author. Journal compilation © 2008 British Educational Communications and Technology Agency. hundreds of demonstrations, clever games and interesting segments, but pitifully few integrated, full courses or even modules of instruction. This pot-pourri may be good for demonstrating the range of possible applications of CAI, but it has not had a major effect on the progress of education and training and has not allowed CAI to take its place as a major contributor to the education and training communities. In addition, it puts an incredible load on the CAI student who must continually learn and relearn the ‘rules of the game’ while passing through lessons generated using an infinite variety of approaches and styles. In the empirical approach, intuition is used to formulate hypotheses, which are then used to design experiments and collect data. Instructional methods or techniques which are supported by the data tend to be accepted over those which do not. The empirical approach to institutional development evolved from the programmed instruction industry and emphasises carefully stated objectives, course try-out and revision based on data from testing students. It has some advantages over the more artistic approaches. It does put some of the onus for poor student performance on instructional developers, and it is likely to eventually lead to more effective courseware. However, it too has problems. The emergence of this empirical model was not accompanied by the development of carefully developed strategies or prescriptive models for getting from well-stated objectives to first drafts of the courseware, or from revision data to final versions. Such strategies and models are left to the individual developer with a few exhortations to use small steps, overt responding and feedback. This model is often prohibitively expensive in CAI applications because it results in long linear sequences, the need for several revisions and many of the same problems associated with the artistic approach described earlier. A curriculum-development activity which uses task analysis, behavioural objectives and formative evaluation and revision techniques is still essentially using a cottage industry’ approach when the carefully stated behavioural objectives are turned over to an instructional developer who operates according to an intuitive or artistic model in order to develop his instructional displays. Among the greatest misuses of CAI have been the applications best described as ‘automated textbooks’ and ‘automated programmed instruction sequences’. Seldom has the very special potential of CAI been so grossly ignored as in these applications. In the systematic approach, analytic procedures or methods are used to design, develop and evaluate instruction. These procedures are derived from a combination of refined experience, accumulated research findings and learning theory. The systematic approach is called instructional systems development (ISD) or just simply instructional development. ISD has emerged in response to the pressing need for a more effective, efficient and reliable educational system. It involves things such as task analysis, behavioural objectives, criterion reference testing, individualised instruction, 254 British Journal of Educational Technology Vol 39 No 2 2008 © 2008 The Author. Journal compilation © 2008 British Educational Communications and Technology Agency. formative evaluation and media selection. Although ISD was originally developed in the context of military training, its use has spread to all domains of instruction and it is probably the dominant instructional methodology existing today. Some important characteristics and benefits of ISD are as follows: